Diagnostic Accuracy of Post-Neoadjuvant Treatment Restaging Magnetic Resonance Imaging in Rectal Cancer in the Era of Total Neoadjuvant Treatment

Abstract

Introduction

The diagnostic accuracy of a restaging magnetic resonance imaging (MRI) scan after neoadjuvant treatment (NAT), especially total neoadjuvant treatment (TNT), is not well established. We aimed to study the diagnostic accuracy of the restaging MRI after NAT including TNT.

Methods

This was a retrospective study conducted at a tertiary cancer center of patients undergoing radical surgery after various types of NAT for rectal cancer between 2020 and 2023. The patients were grouped according to the type of NAT received. The findings on the restaging MRI were correlated with the gold standard of the histopathology report of the resected specimen to estimate its diagnostic accuracy overall as well as for the different types of NAT.

Results

Among the 131 patients included in this study, 68% received TNT. The restaging MRI showed an overall accuracy of 74.4% for T-stage and 60.1% for N-stage and overestimated the T- and N-stage in 20.6 and 12.2% of patients and underestimated them in 41.2 and 47.3% patients, respectively. MRI had a higher accuracy for T-stage among patients who did not receive TNT compared with those who received it (79.6 vs. 75.2%) but not for N-stage or CRM involvement. The overall sensitivity, specificity, positive predictive value, negative predictive value (NPV), and accuracy of the restaging MRI for CRM involvement were 74.1, 73, 26.8, 95.5, and 74%, respectively.

Conclusion

We found a moderate level of accuracy for posttreatment MRI in predicting the pathological stage and a high NPV in determining an involved CRM. Use of TNT reduced the diagnostic accuracy of restaging MRI for T-stage but not for N-stage or predicting CRM involvement. Further studies are required to assess the diagnostic accuracy of posttreatment MRI after different regimens of NAT that are currently recommended in rectal cancer.

Keywords

rectal cancer - total neoadjuvant treatment - magnetic resonance imaging - diagnostic accuracy - restaging

Introduction

High-resolution magnetic resonance imaging (MRI) is the recommended modality for primary local staging in rectal cancer.[1] [2] MRI provides essential information on T staging, mesorectal nodes, involvement of the mesorectal fascia (MRF), tumor deposits, extramural venous invasion (EMVI), and lateral pelvic nodes, which play a critical role in guiding clinicians in tailoring neoadjuvant treatments (NATs) in patients with rectal cancer. The primary objective of NAT is to achieve a negative circumferential resection margin (CRM), a known surrogate marker for both local recurrence and distant metastasis.[3] [4] Reported accuracy of primary MRI for T and N staging ranges from 31 to 100% and 39 to 95%, respectively.[5] Additionally, MRI predicts CRM involvement with high accuracy, which helps determine the need for NAT and also guides the surgeon in planning the most appropriate resection plane.[6] [7] [8] Although the role of MRI in the initial evaluation of rectal cancer is well established, the diagnostic utility of a restaging MRI in assessing the response to NAT is less studied especially when a variety of neoadjuvant therapy strategies including total neoadjuvant treatment (TNT) are currently recommended for use and when subsequent treatment is being tailored according to the response to NAT, ranging from a watch-and-wait strategy to beyond TME approach.[1] While the Tumor Regression Grade (TRG) as assessed by MRI is a recognized prognostic factor for disease-free survival and overall survival, its accuracy in predicting pathological TRG remains suboptimal.[9] [10] The objective of this study was to ascertain the accuracy of post-NAT restaging MRI in predicting the pathological stage and CRM involvement in patients undergoing radical surgery for rectal cancer.

Materials and Methods

This was a retrospective study conducted at a tertiary cancer center. Information was retrieved from a prospectively maintained database of all the patients diagnosed to have locally advanced rectal cancer from the year 2020 till 2023. For the purpose of this study, all consecutive newly diagnosed patients with nonmetastatic locally advanced adenocarcinoma of the rectum who underwent NAT followed by total mesorectal excision (TME) or partial mesorectal excision were included. Patients with other histology, those who did not undergo surgery either because of progression during NAT or placed under a watch-and-wait approach due to complete clinical response and those undergoing palliative surgery were excluded. All the patients underwent MRI of rectum with liver screening along with high-resolution computed tomography of thorax. The MRI parameters have been provided in [Supplementary Material S1] (available in the online version only). MRF involvement on the MRI was defined as tumor ≤1 mm distance from MRF (the primary tumor, radiologically suspicious node, EMVI, or tumor deposit). After completion of workup, all the cases were discussed in the multidisciplinary team meeting. Based on the high-risk features as determined by the MRI, different types of NAT were administered to the patients. For this study, patients were subdivided into six groups based on the type of NAT administered. Group I received long-course chemoradiation (LC-CTRT), Group II received short-course radiation (SCRT) followed by delayed surgery, Group III received SCRT followed by consolidation chemotherapy, Group IV received LC-CTRT followed by consolidation chemotherapy, Group V received induction chemotherapy followed by LC-CTRT, and Group VI received induction chemotherapy followed by SCRT. A response assessment MRI was done 4 to 6 weeks after completion of NAT. While the T-stage was reported taking into consideration both the T2-weighted and the diffusion-weighted sequences. Both the primary staging and restaging MRI were performed in a single institute and reported by a single radiologist who was experienced in reporting rectal cancer MRI. Subsequently, TME was performed 6 to 10 weeks after completing NAT. The histopathological exam was reported by an experienced oncopathologist as per the Royal College of Pathologist guidelines. The CRM was considered involved if the histopathological examination of the resected specimen showed a tumor ≤1 mm from the circumferential margin. The number of patients who had involved CRM and the type of NAT administered in them were recorded. The diagnostic accuracy of the restaging MRI was determined by comparing it with the findings on histopathological examination of the resected specimen, which was considered as the gold standard. Two-by-two tables were used to determine the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and the accuracy of the restaging MRI. Concordance was defined as the agreement between restaging MRI stage and the pathological stage in the resected specimen and calculated as the proportion of MRI reports that agreed or disagreed with histopathology. This analysis was done for all patients first and then for patients who received TNT (Groups II–VI) and did not receive TNT (Groups I–II) separately. This study was approved by the institutional ethics committee.

Results

A total of 469 patients diagnosed with rectal cancer were treated in our institution between January 2020 and December 2023. Among these, 131 patients who met the eligibility criteria were included in the study. The demographic and NAT details are given in [Table 1]. Nearly 68% of the patients received TNT. The most commonly administered NAT was short-course radiation followed by consolidation chemotherapy. The overall median time interval between the completion of NAT and surgery was 7.4 weeks (range: 3–22.5 weeks), while the overall median interval between restaging MRI and surgery was 2.65 weeks (range: 0.28–9.85 weeks). The sensitivity, specificity, and accuracy of restaging NAT MRI in predicting the pathological stage, CRM involvement, and tumor regression grade are mentioned in [Table 2]. The overall concordances for T-stage ([Fig. 1]) and N-stage were 38.1 and 40.4%, respectively. The restaging MRI overestimated the T-stage in 20.6% of the patients and underestimated ([Fig. 2]) in 41.2%, whereas the nodal stage was overestimated in 12.2% and underestimated in 47.3% of patients. The sensitivity of restaging MRI for T-stage was 38.9 versus 45.8% (p = 0.56), specificity was 74.6 versus 83.6% (p = 0.34), and accuracy was 75.2 versus 79.6% (p = 0.74) in patients receiving TNT when compared with those not receiving it. However, the accuracy for N-stage was not different between the groups (TNT: 60.2% vs. non-TNT: 60%; p = 1.0).

Table 1 Patient characteristics (n = 131)

Median age in years (range)	54 (22–80)
Gender
Male  Female	87 (66.4%) 44 (33.5%)
Histopathology
Adenocarcinoma  Mucinous carcinoma  Signet cell carcinoma	118 (90.1%) 10 (7.6%) 3 (2.2%)
Types of neoadjuvant treatment (NAT)
Group 1—Long-course chemoradiation (LC-CTRT)  Group 2—Short-course radiation (SCRT) f/b delayed surgery  Group 3—SCRT f/b consolidation chemotherapy  Group 4—LC-CTRT f/b consolidation chemotherapy  Group 5—Neoadjuvant chemotherapy f/b LC-CTRT  Group 6—Neoadjuvant chemotherapy f/b SCRT	28 (21.3%) 14 (10.6%) 49 (37.4%) 25 (19%) 14 (10.6%) 1 (0.76%)
Median time between completion of NAT and surgery in weeks (range)
Group 1—Long-course chemoradiation (LC-CTRT)  Group 2—Short-course radiation (SCRT) f/b delayed surgery  Group 3—SCRT f/b consolidation chemotherapy  Group 4—LC-CTRT f/b consolidation chemotherapy  Group 5—Neoadjuvant chemotherapy f/b LC-CTRT  Group 6—Neoadjuvant chemotherapy f/b SCRT	9.5 (6.5–22.5) 8.2 (6.5–20) 6 (3.4–14) 6.7 (3–30.1) 10.1 (9–12) 6.2
Median time between restaging MRI and surgery in weeks (range)
Group 1—Long-course chemoradiation (LC-CTRT)  Group 2—Short-course radiation (SCRT) f/b delayed surgery  Group 3—SCRT f/b consolidation chemotherapy  Group 4—LC-CTRT f/b consolidation chemotherapy  Group 5—Neoadjuvant chemotherapy f/b LC-CTRT  Group 6—Neoadjuvant chemotherapy f/b SCRT	2.78 (0.57–6.85) 2.14 (0.57–5.71) 3 (0.42–9) 3.57 (1.28–9.85) 2.71 (0.28–5.28) 1.7
Upfront MRI findings
T-stage
T0  T1  T2  T3  T4	0 0 1 (0.76%) 72 (54.96) 58 (44.27%)
N-stage
N0  N1  N2	1 (0.76%) 27 (20.6%) 103 (78.62%)
Mesorectal fascia status
Involved  Free	55 (41.98%) 76 (58.01%)
Pathological response to NAT
Complete  Incomplete	23 (17.5%) 108 (82.4%)

Table 2 Comparison of restaging MRI stage with pathological findings

	ypT stage (no.)					Sensitivity	Specificity	Accuracy
	ypT0	ypT1	ypT2	ypT3	ypT4
ymrT stage (no.)
T0 (7)  T1 (8)  T2 (37)  T3 (54)  T4 (25)	4 1 9 8 2	0 1 1 3 0	2 4 11 12 5	1 2 16 29 13	0 0 0 2 5	16.6 20.0 30.5 55.7 71.4	97.1 92.9 72.6 63.2 83.8	81.6 91.6 61.0 54.9 83.2
Overall (%)						38.84	81.92	74.46
	ypN stage (no.)					Sensitivity	Specificity	Accuracy
	ypN0	ypN1		ypN2
ymrN stage (no.)
N0 (39)  N1 (71)  N2 (21)	27 45 10	11 22 7		1 4 4		32.9 55.0 44.4	75.5 46.1 86.0	48.8 48.5 83.2
Overall (%)						44.1	69.2	60.1
	CRM status (no.)			Sensitivity	Specificity	Accuracy	PPV	NPV
	Involved	Free
MRF status (no.)
Involved (41) Free (90)	11 4	30 86		73.0 74.1	74.1 73.0	74.0 74.0	26.8 95.5	95.5 26.8

Abbreviations: CRM, circumferential resection margin; MRF, mesorectal fascia; NPV, negative predictive value; PPV, positive predictive value.

Fig. 1 Example of concordance between restaging MRI and pathological stage. Baseline MRI showing mrT3b disease (A). Restaging MRI demonstrating ymrT0 (B), confirmed by histopathological examination showing ypT0 (C).

Fig. 2 Example of discordance between restaging MRI and pathological stage. Baseline MRI showing mrT2 disease (A). Restaging MRI demonstrating ymrT0 (B), while histopathological examination revealed ypT2 disease (C).

The overall sensitivity, specificity, PPV, NPV, and accuracy of the restaging MRI for CRM involvement were 74.1, 73, 26.8, 95.5, and 74%, respectively. The NPV for CRM involvement was equally high in patients receiving TNT and those not receiving TNT (95.9 vs. 96.5%, respectively).

A pathological complete response (pCR) was observed in 17.5% of all patients. The highest pCR rate (24.4%) was seen in Group III followed by Group IV (24%), with MRI showing a 100% PPV and specificity for pCR in both these groups.

Discussion

In this retrospective study, restaging MRI after NAT for rectal cancer showed moderate accuracy for T-stage and N-stage compared with the gold standard histopathology report. The MRI overestimated the T-stage in a fifth of patients and underestimated N-stage in nearly half of them. We also observed a high NPV and a high PPV of restaging MRI in predicting CRM involvement and identifying a complete pathological response, respectively.

The accuracy of primary staging MRI in rectal cancer is well established. However, the diagnostic accuracy of restaging MRI is debatable.[11] The fibrosis and edema as well as acellular mucin pools induced by NAT can limit the interpretation of a restaging MRI. The diagnostic accuracy of restaging MRI is also influenced by the type of NAT (TNT vs. non-TNT), type of radiation (SCRT vs. LC-CTRT), and the time interval between the completion of NAT and the MRI and the surgery. The longer interval between NAT and surgery in TNT and the aggressive NAT in TNT can both lead to severe fibrosis and desmoplasia that can further limit the accuracy of restaging MRI.[12] A recent meta-analysis of 33 studies reported a pooled concordance rate of 63.9 and 60.9% between restaging MRI and T- and N-stage, respectively. The sensitivity and specificity of restaging MRI for T-stage (and N-stage seen in this study) were lower than those reported in earlier studies (87 and 70–71% for T-stage and 77 and 70–71% for N-stage) using only neoadjuvant chemoradiation.[13] [14] This could be due to the inclusion of patients receiving TNT in our study. This is demonstrated by the fact that we observed a higher accuracy of ymrT-stage in patients who did not receive TNT compared with those who did. Further, the observed accuracy of restaging MRI for T- and N-stage in our study (74.4 and 60.1%, respectively) is similar to that reported by Gefen et al. (70.6 and 66.4%, respectively) in a similar population of patients where nearly half of the patients received TNT.[15] It has also been reported earlier that a longer interval between the restaging MRI and completion of LC-CRT increased the concordance between the MRI and pathological stages.[16]

A positive CRM after surgery for rectal cancer is associated with a poor prognosis.[17] While the status of MRF involvement on a restaging MRI helps the surgeon to plan the surgery, predicting an involved CRM on MRI after NAT is difficult due to the persistent hypointense fibrotic change remaining at the initial tumor area.[18] While the sensitivity and specificity of restaging MRI for CRM involvement observed in our study were lower than previously reported values of 76 to 87.5% and 85.7%, respectively[14] [19] in studies using LC-CRT, it is higher than that reported in studies that included patients receiving TNT.[15] Although the high NPV of restaging MRI for predicting CRM involvement that we observed compares with published literature, the PPV of MRI for CRM involvement in our study is lower than the 44 to 57% reported by other studies.[13] [19] This may be attributed to the fact that, when an MRI predicted a positive margin, the surgeon would have performed beyond TME approach during surgery, thereby resulting in a negative CRM on histopathological examination. A low PPV for CRM status on the restaging MRI results in overestimation of CRM involvement. It has been previously reported that restaging MRI can overestimate CRM involvement in ∼30% of patients and CRM proximity by 0.4 cm on average.[12]

The PPV of MRI in identifying a pCR was 100% in our cohort, but due to the small number of patients who achieved pCR, it is difficult to draw any conclusion pertaining to the efficacy of MRI in predicting pCR. Another study reported a PPV of complete clinical response of 73% for identifying pCR.[20] We also found that 16% of patients with an incomplete response (mrTRG: 2–4) had a pCR on resected specimen. This is because the mrTRG does not take into account the nodal status and diffusion-weighted imaging (DWI). Therefore, incorporation of DWI in assessing complete response may help better predict a complete response.[21]

Patients in Group IV (TNT with consolidation chemotherapy) achieved 24% pCR with the restaging MRI showing a 100% PPV and 80% NPV in identifying pCR. This assumes practical significance because this is the preferred NAT in an intentional watch-and-wait approach in patients wanting organ preservation (reference NCCN guidelines).

The strength of our study is that it is a fairly large cohort of 131 patients with minimal variability in imaging and pathological evaluation. Although many previous studies have reported on the accuracy of restaging MRI, TNT was not used in most of these. Our study which included patients receiving TNT, therefore, adds to the growing literature on the accuracy of restaging MRI after TNT. This study is limited by the fact that it is a single-center, retrospective study. The small number of patients receiving different types of NAT makes it difficult to compare the accuracy of restaging MRI between these different treatment regimes. We have not tested for interobserver variability which may limit the generalizability of our results.

Conclusion

In this study of patients receiving different NAT regimens, we observed a moderate level of accuracy for posttreatment MRI in predicting the pathological stage and a high NPV in determining an involved CRM. The use of TNT reduced the diagnostic accuracy of restaging MRI for T-stage but not for N-stage or predicting CRM involvement. Further studies are required to assess the diagnostic accuracy of posttreatment MRI after different regimens of NAT that are currently recommended in rectal cancer.

Conflict of Interest

None declared.

• References

• 1 National Comprehensive Cancer Network. Clinical practice guidelines in oncology (NCCN Guidelines): rectal cancer—version 4.2024. Published on August 22, 2024. Accessed November 10, 2024 at: https://www.nccn.org/professionals/physician_gls/pdf/rectal.pdf
• 2 Glynne-Jones R, Wyrwicz L, Tiret E. et al; ESMO Guidelines Committee. Rectal cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2017; 28 (Suppl. 04) iv22-iv40 (Erratum in: Ann Oncol. 2018 Oct 1;29(Suppl 4):iv263)
• 3 Nagtegaal ID, Quirke P. What is the role for the circumferential margin in the modern treatment of rectal cancer?. J Clin Oncol 2008; 26 (02) 303-312
• 4 Adam IJ, Mohamdee MO, Martin IG. et al. Role of circumferential margin involvement in the local recurrence of rectal cancer. Lancet 1994; 344 (8924) 707-711
• 5 Mari G, Crippa J, Montroni I. et al. MRI-pathology agreement in rectal cancer: real-world data from a prospective rectal cancer registry. Chirurgia (Bucur) 2021; 116 (05) 583-590
• 6 Brown G, Radcliffe AG, Newcombe RG, Dallimore NS, Bourne MW, Williams GT. Preoperative assessment of prognostic factors in rectal cancer using high-resolution magnetic resonance imaging. Br J Surg 2003; 90 (03) 355-364
• 7 Brown G, Richards CJ, Newcombe RG. et al. Rectal carcinoma: thin-section MR imaging for staging in 28 patients. Radiology 1999; 211 (01) 215-222
• 8 Blomqvist L, Rubio C, Holm T, Machado M, Hindmarsh T. Rectal adenocarcinoma: assessment of tumour involvement of the lateral resection margin by MRI of resected specimen. Br J Radiol 1999; 72 (853) 18-23
• 9 Patel UB, Taylor F, Blomqvist L. et al. Magnetic resonance imaging-detected tumor response for locally advanced rectal cancer predicts survival outcomes: MERCURY experience. J Clin Oncol 2011; 29 (28) 3753-3760
• 10 Sclafani F, Brown G, Cunningham D. et al. Comparison between MRI and pathology in the assessment of tumour regression grade in rectal cancer. Br J Cancer 2017; 117 (10) 1478-1485
• 11 Hanly AM, Ryan EM, Rogers AC, McNamara DA, Madoff RD, Winter DC. MERRION Study Group. Multicenter evaluation of rectal cancer ReImaging pOst neoadjuvant (MERRION) therapy. Ann Surg 2014; 259 (04) 723-727
• 12 Yuval JB, Thompson HM, Firat C. et al. MRI at restaging after neoadjuvant therapy for rectal cancer overestimates circumferential resection margin proximity as determined by comparison with whole-mount pathology. Dis Colon Rectum 2022; 65 (04) 489-496
• 13 Wei MZ, Zhao ZH, Wang JY. The diagnostic accuracy of magnetic resonance imaging in restaging of rectal cancer after preoperative chemoradiotherapy: a meta-analysis and systematic review. J Comput Assist Tomogr 2020; 44 (01) 102-110
• 14 Moreno CC, Sullivan PS, Mittal PK. MRI evaluation of rectal cancer: staging and restaging. Curr Probl Diagn Radiol 2017; 46 (03) 234-241
• 15 Gefen R, Garoufalia Z, Horesh N. et al. How reliable is restaging MRI after neoadjuvant therapy in rectal cancer?. Colorectal Dis 2023; 25 (08) 1631-1637
• 16 Aker M, Boone D, Chandramohan A, Sizer B, Motson R, Arulampalam T. Diagnostic accuracy of MRI in assessing tumor regression and identifying complete response in patients with locally advanced rectal cancer after neoadjuvant treatment. Abdom Radiol (NY) 2018; 43 (12) 3213-3219
• 17 Kulkarni T, Gollins S, Maw A, Hobson P, Byrne R, Widdowson D. Magnetic resonance imaging in rectal cancer downstaged using neoadjuvant chemoradiation: accuracy of prediction of tumour stage and circumferential resection margin status. Colorectal Dis 2008; 10 (05) 479-489
• 18 Seo N, Kim H, Cho MS, Lim JS. Response assessment with MRI after chemoradiotherapy in rectal cancer: current evidences. Korean J Radiol 2019; 20 (07) 1003-1018
• 19 Jia X, Zhang Y, Wang Y. et al. MRI for restaging locally advanced rectal cancer: detailed analysis of discrepancies with the pathologic reference standard. AJR Am J Roentgenol 2019; 213 (05) 1081-1090
• 20 Lambregts DM, Vandecaveye V, Barbaro B. et al. Diffusion-weighted MRI for selection of complete responders after chemoradiation for locally advanced rectal cancer: a multicenter study. Ann Surg Oncol 2011; 18 (08) 2224-2231
• 21 Chandramohan A, Siddiqi UM, Mittal R. et al. Diffusion weighted imaging improves diagnostic ability of MRI for determining complete response to neoadjuvant therapy in locally advanced rectal cancer. Eur J Radiol Open 2020; 7: 100223

Publication History

Article published online:
30 May 2025